2020/07/03 by Max Biggs, Wei Sun, Biggs, Max +4 · 1 citation
Computer Science · Decision Sciences · #Advanced Bandit Algorithms Research #Applications (stat.AP) #Data Stream Mining Techniques #Explainable Artificial Intelligence (XAI) #FOS: Computer and information sciences #Forecasting Techniques and Applications #Machine Learning (cs.LG) #Machine Learning (stat.ML)
paper · pdf · doi:10.48550/arxiv.2007.01903
openalex publication_date 2020/07/03 · openalex created_date 2022/07/26 · openalex updated_date 2026/07/28
Data-driven pricing strategies are becoming increasingly common, where\ncustomers are offered a personalized price based on features that are\npredictive of their valuation of a product. It is desirable for this pricing\npolicy to be simple and interpretable, so it can be verified, checked for\nfairness, and easily implemented. However, efforts to incorporate machine\nlearning into a pricing framework often lead to complex pricing policies which\nare not interpretable, resulting in slow adoption in practice. We present a\ncustomized, prescriptive tree-based algorithm that distills knowledge from a\ncomplex black-box machine learning algorithm, segments customers with similar\nvaluations and prescribes prices in such a way that maximizes revenue while\nmaintaining interpretability. We quantify the regret of a resulting policy and\ndemonstrate its efficacy in applications with both synthetic and real-world\ndatasets.\n